remove outliers form timeseries

11 views (last 30 days)
Gaetano Pavone
Gaetano Pavone on 11 Nov 2019
Edited: Gaetano Pavone on 12 Nov 2019
I have an original timeseries ts and I would like to apply a filter for removing the outliers. In particular I define a percentile criteria for filtering the original ts:
[B,TF]=rmoutliers(ts.Data,'percentiles',[5 95]);
Such operation returns two objects: a filtered ts.Data, that is B, and the logical vector Tf in which if the element is outlier the corresponding element is equal to 1, otherwise is equal to 0.
Now my question is:
how to build a timeseries with the filtered data B and the corresponding datetime vectors?

Accepted Answer

Daniel M
Daniel M on 11 Nov 2019
If timevector is the name of your datetime vector, then you can mask out the outliers using
timevector_good = timevector(~TF); % this is the "good" data points
  1 Comment
Adam Danz
Adam Danz on 11 Nov 2019
You can use the TF logical vector to identify non-outlier indices.
My answer moved here as a comment to supplement Daniel's answer.

Sign in to comment.

More Answers (0)





Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!